Is no preamp the best preamp of all?

I'm sorry, but this analogy is even less illuminating than the last one. I'm not afraid of a bit of technical language if you want to use it.

"those least significant bits contain detail."

Of course, I never disputed that, though it is disputable how much detail can be heard at -80 or -90 dB down. What I'm am disputing is the idea that turning down an analog volume control does not also lead to a loss of information in the signal that reaches one's ears.

OK, after a bit of thought I think I see what you are trying to say with the analogy. It helps me to think of it in terms of arithmetic. If you start out with 16-bit audio you have 2**16 different peak-to-peak voltages available. If you reduce the volume by 12 dB you then only have about 2**14 available voltage steps, which is perhaps where the don't-go-below 75% maxim comes from (75% being about a 12.5 dB cut in .5 dB steps.) So don't go below 75% or you playback sounds like an old 14-bit CD player!

But most digital volume controls these days are 24-bit or 32-bit. The Squeezebox outputs 24-bit words, and shifts 16-bit audio to the most-significant bits of each word. This doesn't magically give you 24-bit resolution, but it does mean that with a 24-bit DAC you retain at least 16-bit resolution even with a quite large attenuation (theoretically 8 bits or about 48 dB; but that's nearly the full range of the volume control). Even with 24-bit files, I wouldn't worry until the attenuation was more than 24 dB (or about 4 bits) since you can't really expect to get more than about 20 bits of resolution anyway.

08-26-11: Daverz
What does 75-85% translate to in terms of dB?

I'm aware that I'm losing information when I use digital attenuation, but how does this translate into what one hears that is different that what one hears when an analog volume is lowered. Saying that one "loses resolution" is not really telling me anything as I don't know how a "loss of resolution" sounds different from just a lower volume.

it is not meaningful to just cite a figure like "75%-85% because you need more information. as a general matter, each 6dB reduction corresponds to a 1/2 reduction in power (or 3dB reduction in voltage). this means that you are losing 1 bit of resolution for each 6dB. redbook cd stores audio with 16 bits of resolution/channel (but the actual resolution, as a practical matter, is more like 14 bits). so what you need to know is how many bits of resolution is used by the audio processor in your cd player. i have a wadia 381, which operates with 21 or 22 bits of resolution. each step in the wadia digital volume control reduces the (power) output by 1 dB. so, i can reduce the digital volume to around 65% to 75% before i start to face the potential of losing significant bits of audio information.

my personal belief is that these are the considerations for deciding whether to go through a pre-amplifier or direct:

1)some people like the way that the pre-amplifier colors the sound from the cd player. such people should go through the pre-amplifier instead of going direct;

2)when i want to listen a low volume levels (like late at night), i tend to go through the preamplifier. this gives you a lot of latitude in how you manage volume: there can be drawbacks to turning the volume level too low on a pre-amplifier, so you set the volume level on the pre-amplifier to be in the range you desire and then adjust the digital volume to bring the volume the rest of the way down;

3)there are some recording (like some classical recordings) where the recording level is much lower than that of many popular recordings. in such cases, you would need to go through the pre-amplifier to get sufficient gain to achieve the volume level that you desire.

I believe the wide range of reported successes and failures of passive attenuation is based on the many variations in matching sources and amplifiers. Arthur Salvatore writes about this on his website (Link:http://www.high-endaudio.com/RC-Linestages.html#Int) and summarizes with;

In short, if you need an active line stage because your source is not up to the task of driving the amplifier(s), then...

Any good active line stage, from any era, will improve the sonics in some obvious and clear manner.

Alternatively, if your source is up to the task of driving your amp(s), then...

No active line stage, no matter how good it is, will ever equal the sonics of your direct connection (or an equivalent passive).

Regarding digital volume control, here is an interesting paper by Daniel Weiss;
http://www.weiss-highend.ch/computerplayback/Digital_Level_Control.pdf

Arthur Salvatore has updated and revised his opinion since he auditioned ans subsequently purchased the active Coincident Statemnet Linestage a few months ago. He now believes for the vast majority(with rare exception) a superior active is preferred.